Using block-based multiparameter representation to detect tumor features on T2-weighted brain MRI images

Area of Science:

• Medical Imaging Analysis
• Computer-Aided Diagnosis

Background:

• Accurate tumor detection is crucial for medical decision support systems.
• Existing methods may lack efficiency or simplicity in image analysis.

Purpose of the Study:

• To present an analytical method for detecting tumors or lesions in digitized medical images.
• To develop a computationally efficient and conceptually simple tumor detection technique.

Main Methods:

• A novel method utilizing edge (E), gray (G), and contrast (H) parameters was developed.
• The method analyzes Visible Human Dataset (VHD) brain images by dividing them into fixed-size blocks (templates).
• EGH parameters were calculated for feature blocks to identify and mark abnormalities.

Main Results:

• The proposed technique demonstrated effectiveness in detecting tumor blocks across various medical image datasets.
• Experimental results confirm the method's ability to handle different time intervals, brain diseases, and multiple image slices.
• The system achieved conceptual simplicity and computational efficiency in tumor detection.

Conclusions:

• The study successfully developed and tested a prototype system for tumor detection.
• The presented method offers a promising approach for enhancing medical decision support systems through efficient image analysis.